Cosmology and verifiability

George R. Ellis

pp. 93-113

in: Robert S. Cohen, Mark W. Wartofsky (eds), Physical sciences and history of physics, Berlin, Springer, 1984

Abstract

Relativistic Cosmology aims to determine the structure of the Universe from a fusion of the results of astronomical observations with knowledge derived from local physical experiments. The problem of determining this structure¹ is centred on the fact that there is only one universe to be observed, and that we effectively can only observe it from one space-time point. Because it is a unique object, we cannot infer its probable nature by comparing it with similar objects; and (on the scale we are considering) we are unable to choose the time or position from which we view it. Our predicament is analogous to that of a premaritime man living on a small island in an ocean, who observes around him a host of other small islands apparently scattered at random on a seemingly limitless sea. Unable to move from his island, his theory of the world in which he lives can only be based on this partial view.